Message-ID: <15cc1c95baf.marcocadeddu@tin.it>
Date: Mon, 19 Jun 2017 21:16:57 +0200 (CEST)
From: "marcocadeddu@tin.it" <marcocadeddu@tin.it>
To:  <rsgb_lf_group@blacksheep.org>
Mime-Version: 1.0
Subject: R: Re: Re: R: Re: LF: Re: I: Fw: For today the FETs survived...
Content-Type: text/plain;charset="UTF-8"
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Content-Transfer-Encoding: 8bit
Status: O

...me again on the half bridge and guard circuit ;-)
I found an interesting reading above this circuit wrote from .. G4JNT: http://www.g4jnt.com/OverloadProt.pdf
so hopefully there is somebody here around that can answer my doubts ;-
)

But before explain my new mumbling, better to start from the 
beginning..
After finally had the amplifier running my worse problem for testing 
was the dummy load: my Cantenna oil cooled is not the best to run for 
long tests with more than 500W, so before to connect to the antenna (I 
do hope it can cope with it..) I returned to the guard circuit.
Of course, as usual I took the wrong way monitoring the current 
changing the load, so running the ampli at low power I started to 
change the numer of turns in the link coil minimizing the return 
current with load 50 ohms.
The start point was a 4:1 ratio as suggested between tank coil and link 
coil thus being the tank coil a 3 layers coil of 86 turns, an available 
wire made be happy with a link coil of 18 turns.
When arrived at no current the link coil had only 9 turns.. I made some 
test with short or open circuit at 50Vdc and the amplifier survived :-) 
but the return I guard was not so impressive: just a fraction of normal 
Idd (normal Idd 1A@50V, with shorted antenna Idd 0,36A and Ig 0,5A).

The started the tests with direct rectification of the mains and was 
time to test the guard circuit..
I had no time to record some datas.. a cap of the tank started to 
arch.. I quickly diconnected all but one of the caps was half 
vaporized! I replaced it and all (but the guard circuit) was working 
again.
At the same time I was kept from the desire of a better tank coil.. the 
first one was coiled on a paper tube placed on 2 pvc saddles and kept 
in place from a wrap..
So after a visit to the hobby center I came back with a nice, orange 
pvc pipe 100 diameter and 25m o 2.5mm² electric wire.
I prepared a new, neat tank coil with 2 nice supports. As usual is 
better to cut the excess of turns but I wanted to see what could happen 
;-)

So the "old" tank coil had 3 layers of turns (89T in total) starting 
from 65mm of diameter and 140mmlong; the measured inductance was 223uH 
with 28pF of stray capacitance.
It was resonated at 136.2 kHz with a combination of 8x3.3nF+2x1000pF 
(abt). The 3dB BW of this tank is 23.7 kHz.
The new coil is coiled on the pvc form 100mm dia, 110 mm long wit a 
total of 62 turns. It is much better in apparence but.. is 285uH with 
stray capacitance of 175pF!
It was resonated at 136.8 kHz with a combination of 6x3.3 nF (abt 
4950pF) and shows a 3dB BW of 21.5 kHz

And now we arrive at the point.. The power supply is the same, the PA 
is the same, the output xfmr is the same (15T/19T on 196 mm² core), the 
load is the same but:

Vdc  Idd  Pout
308  1.84 585W
320  1.56 465W

I assume that the difference in the Vdc is due to a lower Idd (with no 
load in both case the supply is 320Vdc)
With the old tank coil the XL of the coil make the tank circuit more in 
the area of Q6, altough the 3dB BW suggests a Q of 5.7 for the old tank 
and of 6.4 for the new one, but

before I start to cut turns, is that enough to justify almost 100W less 
in the output?

Thank for sharing your thoughts

73 Marco, IK1HSS

----Messaggio originale----
Da: andy.g4jnt@gmail.com
Data: 4-giu-2017 18.42
A: <rsgb_lf_group@blacksheep.org>
Ogg: Re: Re: R: Re: LF: Re: I: Fw: For today the FETs survived...

Yes.
You need to adjust the link so current is driven back into the supply 
when
RF amps exceed the maximum that you have designed for.   Up to that 
point
there should be no guard current as the induced voltage in the link, 
when
rectified , is less than the supply.

If the link coupling were to be perfect, ie it behaved as a 
transformer,
then at full power into your designed load the voltage across the tank 
coil
would be Q * Vdc * 4/pi  * SQRT(2)    (since you calculated Rload this
way).    The SQRT terms comes from the DC being the peak

So the turns ratio (for perfect coupling) if made equal equal to 1 : Q 
* 4
/ pi * SQRT(2) ,will start to generate load current just as maximum 
output
is reached.

But things aren't perfect; the mutual coupling is never 1 in an air 
wound
coil (mine is about 0.6), the rectification doesn't   give exactly peak
volts, so you have to experiment, changing the link turns until you get 
the
limiting you want.   Fortunately it scales with DC volts, so guard 
circuit
testing can be done at low supply / power out.

'jnt

On 4 June 2017 at 16:47, marcocadeddu@tin.it <marcocadeddu@tin.it> 
wrote:

> heyheyhey..
> now I'm testing with the new coil still at 50Vdc ;-)
> ops I was forgetting: without dissipator :-)) just with the mounting
> alu block it runs at room temp, maybe it is worth later at full 
power!
> The
> picture of Andy is not in thema: is the consequence of qro ampli on
> undersize external loading coil :-)
> By the way when this ampli will finally run on the antenna I need to 
re-
> size the variometer because it warms already with 250W..
>
> Also the new tank coil with 2,5mm² wire after 1 hour has a bit of
> fever.. but is not warm as the former.
>
> Just a quick question Andy: speaking of the guard circuit, if I read
> Iguard whithout mismatch is correct the assumption that the link has
> too many turns?
>
> Alan: I'm vintage.. and I'm sure I have a variac in the basement :-)
> thanks for the hint ;-)
>
> Thanks all for the psycological and praxi help !!
> Marco
>
> ----Messaggio originale----
> Da: andy.g4jnt@gmail.com
> Data: 4-giu-2017 17.49
> A: <rsgb_lf_group@blacksheep.org>
> Ogg: Re: R: Re: LF: Re: I: Fw: For today the FETs survived...
>
> Something like this ...
> http://www.g4jnt.com/QRO_LF_DoesThis.JPG> 'jnt :-)
>
>
> On 4 June 2017 at 16:25, Chris Wilson <dead.fets@gmail.
com> wrote:
>
> > Hello Alan,
> >
> >
> > No, what he needs is a video camera, and the courage 
to post the
> > result here for scrutiny, we need a bit of fun now and again ;)
> >
> > Good luck Marco, what's the worst that can happen.... :)
> >
> >
>
>
>
>
>